The latest epidemiological surveys show that people with mental disorders account for about 16 to 17% of the entire population, and have a lifetime prevalence. This means that about 230 million people have suffered from a certain type of mental illness in their lifetime.
In contrast to these 230 million people, there are only less than 50,000 occupational psychologists. At the same time, if you want to cultivate a professional psychological counselor with sufficient professionalism and humanistic care, the high cost of immersion and the large cost of time make the balance of this supply and demand relationship more and more serious.
We can't help but ask, is it possible to use robot psychotherapists to assist, or even replace, some of the traditional psychological diagnosis and treatment work to help more people who are deeply trapped in the psychological vortex, just as traditional medicine replaces human hands for high-precision surgical operations?
Robot psychotherapists, how far are we from being?
In the recently concluded "TCCI Applied Neurotechnology, Artificial Intelligence and Mental Health Forum", Li Chunbo brought his views.
Li Chunbo
Chief physician and doctoral supervisor of Shanghai Mental Health Center
Institute of Psychological and Behavioral Sciences, Shanghai Jiao Tong University
Chinese researcher at the Tianqiao Brain Research Institute (TCCI).
, duration 15:30
Hello everyone, today I will introduce you to you according to the topics collected from past questions, from the perspective of clinical research on artificial intelligence and mental health.
01
Why is mental illness difficult to treat?
The first question is, why is mental illness still so difficult to diagnose?
As you may know, a psychiatrist is still using a very traditional method of diagnosis and treatment in the current process: relying on medical history and interviews.
That is, through an exchange with patients and families, relevant information is collected, and finally summarized to form a subjective judgment. The basis of judgment is mainly some relevant international or national relevant diagnostic criteria, through which the doctor conducts a comprehensive analysis.
At the same time, the doctor will also use some auxiliary tests. Ancillary testing refers to the final diagnosis made by excluding some other organic disease, or other categories of disease.
These diagnoses, including some treatments, are mainly through speech therapy. The so-called speech therapy is to use communication methods such as psychological counseling to judge the performance of some patients.
So far, the exploration of the pathological mechanism of mental disorders is still ongoing, and there is no very consistent opinion or conclusion. Therefore, there is still a long way to go to explore the root causes of mental illness. This requires more people's attention, including the support of the state, such as our country's "brain plan".
02
What is the general level of diagnosis and treatment of mental disorders in China?
As a clinician, the group we serve is primarily a variety of mental disorders. What is the proportion of mental disorders in our country?
With such a large number of patient groups, the number of psychiatrists in our country is very scarce compared to them. However, to reduce this 16 to 17% ratio, it is more necessary to pay more attention to psychological and behavioral problems earlier, to find and solve problems in a timely manner earlier, especially by maintaining and promoting mental health to reduce the entire proportion, which will be possible.
The people we serve are constantly changing. The development of artificial intelligence and the latest hard technology has provided us with new possibilities for related diagnosis and treatment, which is what we call digital therapy.
The so-called digital therapy refers to a method of treating a certain type of mental disorder through a software program. It uses evidence-based evidence to effectively alleviate a current situation of mental disorders in humans.
So can digital therapies be used to evaluate? Can it be used for prediction? And what we're most concerned about, can it be used for treatment?
At the same time, I would like to ask you three more small questions:
Do you think there is a bright future for the digital health industry?
When the technology matures, will you use digital health for your patients?
Are you willing to embrace digital health?
03
How is digital health in the field of mental illness progressing?
I would like to explain the clinical progress of digital health from five aspects.
(1) Real-time detection and tracking of patients
The first is the use of networks or digital technologies for real-time detection, tracking and intervention of emotions or behaviors.
Earlier, a colleague mentioned in the question whether he could use some voice technology to assist in diagnosis and treatment. In fact, in the current research on mental disorders, with the gradual popularization of network and digital technology, the large use of smart phones and wearable devices and other auxiliary tools, for the overall and individualized emotions or behaviors of the population, especially a practical performance in natural situations, has been able to detect and track in real time, and even at the level of intervention has become more feasible.
I would like to give an example of bipolar disorder. Bipolar disorder is a mood disorder disease, which is manifested in a certain period of time, when it is very excited and manic; in another period of time, it is extremely depressed or even negative, and this tendency and even behavior of suicide occurs.
Some researchers mentioned in last year's publication that under the premise of informed consent, they can analyze and predict the change of patients' mood scores within the next seven days by tracking the voice of smart phones of patients with bipolar disorder. Thus, we can predict changes in mood through a feature of speech.
In another article in Nature magazine last year, Li Feifei's team mentioned: How to use sensors in the environment, or even some of these sensors in your home, to collect relevant information to help health care?
Such as voice, which is a feature of your voice, and a frequency of movement in your environment, including a change in the social network, etc., these can become an information variable to form a real-time monitoring and tracking of the prediction model of the variable.
In an article published last year, researchers used smartphone measurements to divide the collected data into two broad categories: one called active data and the other called passive data.
The so-called active data refers to some symptoms of the patient's self-evaluation, such as whether there is anxiety, whether there is a bad mood, whether there is sleep problems, etc., which are self-reported, that is, active data.
Passive data, on the other hand, does not require patients to proactively provide it. Through sensors, including phones and some other types of products, you can know a rhythm of your social interactions, such as the amount of time you spend at home, the number of times you interact with other people, and the distance away from home. For example, there is a depressed person who does not move at home all day and does not communicate with people, which reflects a certain trait of him.
In the study of psychiatric organisms, it was found that through these collected wearable device data, a numerical phenotype can be established, the characteristics of the disease can also be reflected, and a difference between patients and healthy people can be well distinguished.
A recent working study found that among the current wearable devices, there are three variables that are most used and considered effective: heart rate variability, EEG, and skin changes. Some of the other indicators are sensitivity, and the characteristics are not so high.
You can look at a picture on the right. These are some of the sample sizes involved in this review.
Unfortunately, most studies have based these conclusions on data from less than 100 people. The data of a few of them may be hundreds or thousands of people, but why are so many of these studies such a small sample size?
Because you have to have a detailed assessment and diagnosis of each individual, you can determine whether this person is a patient or a temporary fluctuation in mood.
Therefore, in the research of wearable devices, one of the future development directions is how to improve the dimension and accuracy of some samples with higher efficiency.
(2) Psychotherapeutic application
In psychiatric treatment, another scenario in which digital medicine can be applied is psychotherapy.
I just mentioned that it is speech therapy, that is, through the dialogue between the psychotherapist and the client to treat, improve, or remove some of his symptoms.
Psychotherapists themselves need to go through this process to continuously accumulate experience and continue to grow. Therefore, in the process, someone needs to supervise, which is what we call a supervisor. But in fact, there are very few psychotherapists themselves, and supervisors are even more scarce.
In this field, Li Feifei's team collaborated with clinical psychologists abroad to make a preliminary inquiry: researchers recorded the voice in this psychotherapy process and analyzed it automatically through the computer to explore some of the high-risk words.
For example, when the patient mentions some negative related words, "go to the edge of the cliff", "go to the window", "go to the edge of the building", "very uncomfortable", "feel very depressed and even want to cry", etc., it can accurately identify and observe whether the psychotherapist has some responses in this. If the patient mentions these negative meanings and expressions, the therapist does not respond
Then this digital medical system will prompt the therapist: You should pay attention to the next time, you have ten of these appearances, you missed three or four times, as an automatic supervision to assist the therapist.
Of course, this is still being explored, and these need to be completed with the support of many information technologies, including artificial intelligence.
(3) Chatbots
The third area is chatbots, and other emerging web or digital technology applications.
The so-called chatbot is a computer program that simulates human conversation to achieve the purpose of chatting. It has a good convenience, as well as feasibility, popularity, low cost and other characteristics, so it is increasingly used in computerized psychotherapy.
More successful in foreign countries, the use of a wider range of use is a chatbot called Woebot. In fact, it is still in its infancy, and it is all procedural dialogue. There are really some personalized contents, or to be further realized by artificial intelligence.
There are also some app programs that are very common in smartphones to treat depression, and so far there are five models.
The first is that after face-to-face psychotherapy, users can complete some relevant homework and exercises on the APP.
The second is called the ecological immediate intervention mode, which is the information collected through the sensor, such as you have not moved at home all day, or you have not talked to a person today, at which time it will immediately prompt you. "You should go out for an event", "You should go talk to a friend", etc.
The third is self-directed learning and practice mode. Users can also easily open the APP anytime, anywhere with related products.
The fourth is the game mode, that is, through some role-playing games, to combine with some content in real life, and to treat patients' diseases through gamified upgrades.
The fifth is the mutual assistance counseling model, which includes some peer online support, multidisciplinary participation and learning. In the UK National Health System, which is its health insurance system, the recommended mobile app can be prescribed to patients as prescription content.
(4) Virtual reality VR technology
The fourth is to use virtual reality VR technology for psychological counseling and psychotherapy.
As we all know, using VR technology to create an immersive interactive environment can enhance the authenticity and safety of psychotherapy
And the feasibility and convenience will be greatly improved.
For example, in 2018 in the Lancet Journal of Psychiatry, from KCL (King's College London, UK).
In an article published by Professor Tom Craig's team, he conducted a randomized controlled clinical trial to treat the auditory hallucinations of schizophrenia with 50-minute avatar therapy six times a week.
The so-called avatar therapy refers to the establishment of a visual avatar in line with the patient's auditory voice and image, the therapist drives the avatar to dialogue with the patient in real time, switches to the therapist identity according to the patient's on-site response, encourages the patient to talk with the avatar, changes the patient's cognitive mode with the help of real-time dynamic three-person dialogue mode (patient, therapist, avatar), supports the patient to gain initiative and control in the interpersonal relationship with the auditory voice (patient and auditory hallucination content), and ultimately alleviates the symptoms of auditory hallucinations.
For patients, the original uncontrollable auditory hallucinations, through active control, can alleviate the symptoms of patients' auditory hallucinations, so as to achieve the purpose of treatment.
After the end of treatment, this effect can last up to 24 weeks. However, after half a year, it will be closer and closer to the effect of supportive psychological counseling in the control group, but it will still remain in a relatively good state.
Further research in this area focuses on how to achieve long-range effects and how to treat them at any time.
(5) Digital drugs
Finally, there is a word that has been trendy in the past two years: digital drugs.
Digital medicine is a software program that provides an evidence-based approach to treatment. It works the same way as drugs.
The first digital drug to appear internationally last year was Professor Adam Gazzaley from the UCSF (University of California, San Francisco). It is FDA (U.S. Food and Drug Administration) certified. It is primarily suitable for the treatment of the core symptoms of ADHD's attention deficits.
It can be seen in the advertisement, "Play while healing".
This is an advantage of digital medicines, with very convenient possibilities and can be used for future expanded services
An approach for preventive management and rehabilitation.
As a physician, I would say that the most we can see in the clinic is usually the patient, or the high-risk group. How to evaluate them, test, diagnose, then treat, and even to rehabilitation, I hope that through the development of digital products, these can reach a higher level in the future.
In the longer term, how to nip it in the bud from the perspective of primary prevention, from normal individuals to susceptible individuals, I think this is the development discovery that this type of digital product can seek, to play its greater role.
Now that artificial intelligence technology is developing more and more deeply, we can expect that through these digital products, we can complete these psychological tasks at home and at school.
Finally, I would like to say that there is no health without mental health.
I hope that all colleagues will unite in a multidisciplinary manner to pay more attention to mental health. As mentioned in the national "brain plan", it is necessary to concentrate on doing big things.
That's it, thank you!
Excerpted from the WeChat public account